Capacitor substitution

[flinx]

I scavenged a simple 16v 1.6amp DC power supply circuit from a radio. The circuit takes 115v AC in. The 2200uf 25v capacitor on it is just slightly bulged. I have done some researching to find out what the ramifications are of replacing the 2200uf capacitor with say an 1800uf or a 3300uf but could not find anything helpful.

What effect will it have on the output current/voltage?

[mariushm]

If I remember correctly, lower capacity would cause more ripple after the input is rectified and the input voltage may go low enough to cause problems with the switching mosfet/whatever the circuit uses. If that is forced to turn off at certain points you may have more unstable voltage at output.

Too much capacity may overheat the rectifying diodes/bridge rectifier as the pulses of power going into the capacitor will be higher. Usually this is not a problem as the diodes are cheap enough that usually they're bigger than needed so they can take the hits.
1800uF doesn't sound so bad but I wouldn't do it ... I think 3300uF won't be a problem.

Don't you have some store around you to get a capacitor of the proper rating? They're less than 1$ each (or equivalent in your country)...

[budm]

The value of the capacitor will affect the AC ripple voltage riding on top of the DC voltage, it depends on how much current you are going to pull from the power supply, that will affect the amount of AC ripple, so you need to know how well your load will accept the ripple voltage.
learn more here:
http://sound.westhost.com/power-supplies.htm

http://www.electro-tech-online.com/e...ly-design.html

http://waynestegall.com/audio/ripple.htm

[flinx]

The power supply is not regulated. It has a transformer. connected to a small circuit board. there is a bridge rectifier circuit but not an actual bridge rectifier. I was just curious since I don't have any 25v 2200uf capacitors. Actually I don't have any 25v 3300uf either.

[budm]

If you can tell us more of what you are trying to do, what kind of load the power supply will have to deal with, that will be helpful.

[Evil Lurker]

From my understanding with a simple unregulated linear power supply the capacitance of the filter cap(s) = more is better. But like you said too much capacitance could blow the diodes from the inrush current. Me I wouldn't hesitate to put a 3300uf cap in place of the 2200uf and see what happens.

FWIW, I got a linear PSU that I'm working on right now that I'm debating on putting about 64000uf worth of capacitance on.

[flinx]

Originally Posted by budm

If you can tell us more of what you are trying to do, what kind of load the power supply will have to deal with, that will be helpful.

not really trying to do anything. Its just like I said. I scavenged a PS. It has a possibly bad (or going vad) cap. Just wondering what I can substitute, and what would happen if I did.

Since the PS is 16v 1.6 amp. That would be the max load I might use it for. I don't have a specific use in mind at this time.

[jetadm123]

Another possibility would be to connect another cap in parallel with the 1800uf cap. Say you add a 470uf 25V cap in parallel with the 1800uf. That would give you 2270uf.

[flinx]

picture is worth a thousand words.




Yellow is input AC voltage from transformer. Red and Black DC output.

Any idea what the little ceramic disk caps are for in the rectifier circuit? Couldn't find a similar circuit on line.

Oh and should I add a bleeder resistor after the large cap? if so what specs? This thing holds a charge when power and load are removed.

[mariushm]

A capacitor of this size should leak by itself within about 5-10 minutes.

Also, since we're talking of a 16v power supply, you'll only have 16-18v on that capacitor so it's unlikely to give a shock, unless you lick the terminals - the skin has a bit of a resistance.

You can desolder that capacitor and solder two 1000uF instead... if you can route both wires of the capacitors through the holes if they're large enough and mount the capacitors horizontally on the board.
Or, better... carefully drill the pcb and scratch away the insulation from the copper pads and solder the capacitor wires on the copper....

Two capacitors in parallel will have double the capacity so you can just solder additional caps:

caps.jpg

Click picture to zoom it...

PS. This board is not actually the power supply, just to be clear. The AC voltage from the transformer (not 115v AC, around 12-15v AC) comes through the yellow wires, goes through the fuse and then goes through the four diodes at the top. Next, the rectified voltage goes through the two capacitors at the bottom.

The capacitors near the diodes I assume they're used to attenuate the effect of the voltage drop on the diodes (0.4-0.7v) and that very short moment the diodes commute their state.

[flinx]

If I were to drop the voltage down to a more common usable voltage like 12v, how should I do it? how will that effect the Amperage?

I measured the output voltage of the transformer and it is 12v AC.

[mariushm]

The transformer outputs 12V AC, but when it goes through the diodes, it is rectified, so the voltage you get is 12 x √2 = 12x1.41 = 16.92 V DC.

But the diodes have some voltage drop on them, about 0.5-0.8v and since two diodes are conducting at every moment, what you have at output is about.. taking the worst case scenario 16.92v - 1.6v = 15.42v

Now it also depends on the transformer and how much you pull from it... the transformer may be 12v at 0.2A but may be 11v at 2A - the A rating of the transformer matters.
So I'd say you may have 13-16v DC out after rectifying.

Also note that if you have 12v AC 2A transformer, it doesn't mean you get 13-16v 2A DC, but only about 1.6A DC... basically the total VxA = VA (W) remains the same, minus a few losses in the diodes.

If the total Amps on the DC is small (let's say 2-4A) you can use a linear regular to convert that 13-16v to a stable 12v DC.

For example, a chip like this one can do it easily:

http://www.digikey.com/product-detai...223-ND/1616883

This chip takes anything over 12v + about 0.4v and outputs fixed 12v - but it can output a maximum of 3A of power.

They're very easy to use ... this particular chip you connect the first two pins to input (12.5 DC or more) , the third to ground and the fourth pin is the output, 12v DC @ max 3A .. the fifth pin you leave unconnected
You just have to add a capacitor (470uF-1000uF 16v or better is enough) on the 12v DC out and you're done.

If you need more than 3A there are other chips like LT1084 that can do 5A and they're even easier to use with only 3 pins (http://www.digikey.com/product-detai...3PBF-ND/888609) but they're more expensive. (and this particular chip also needs at least 1.3v over 12v to output stable 12v, when you demand 5a from it.. at less than 5a it doesn't need that much over 12v)

But with both of these, you have to keep in mind they convert the excess voltage in heat, so for example if you give them 16V and want 12v out at 3A then you have (16-12) V x 3A = 12W of power wasted as heat, so you also need a good heatsink, something like this for example:

http://www.digikey.com/product-detai...380-ND/1216384
http://www.digikey.com/product-detai...64E-ND/2416494
http://www.digikey.com/product-detai...350-ND/1216357

and you need some thermal paste between this chip and the heatsink (the kind you put on cpu)
You can manage with smaller heatsinks if you have a fan inside blowing air over it.

There's also a question of efficiency, as these chips are about 80-85% efficiency but it already seems to me it gets too complicated for you.

Switching regulators are better, don't lose power as heat like linear regulators, but they're also harder to design at high A values... it's almost easier to just use a cheap regular computer power supply than to build a 12v switching power supply by yourself...

[flinx]

I'm learning a lot here.

On the transformer the amp rating is unknown. The only numbers on it are NC-U4349TM and above that a 01.

[flinx]

On an old LCD power suppy board I found a

KIA78R12API

It looks like it would work but I am not sure how to wire it.

[mariushm]

That one has 4 pins :

IN (DC)
OUT (DC)
GND
ON/OFF (unconnected = ON)

So you connect the + of that board to the IN of this chip, GND to GND pin, and you connect two wires to the OUT (+) and GND (-) and that's your 12v output. The ON/OFF you leave floating in the air, the chip is by default ON.

Note that this chip is rated for maximum 1A, so it will probably turn itself off and on if you pull more than 1A from it.

kia12.png

(minor correction: that's "optional, **should use** if the wires from the input capacitor to it are long, say... more than 5-10cm)

You'll NEED a heatsink for it otherwise it will keep turning itself off and on due to overheating if you use it with more than about 12.5-13v.

[budm]

Learn more, see PDF

[999999999]

I feel this topic is going a bit backwards. Usually, you start out with a project with defined needs, you have a load needing X voltage and Y current and you determine if any of the parts you have are suitable to build a power supply for that. Your power board could be a start to that but what you have on the board is quite basic, it might be more of a hassle to mount a 2nd board than just throw a bridge rectifier, capacitor or two, and a fuse on the load's board, of course if that's reasonably possible.

Also, playing around with LOW voltage circuits is a good way to learn, but when you get to the point where you're ordering parts to build something, the power board and regulator you salvaged are the equivalent of only $1 to $2 worth of parts (depending on your global location), less considering you have to replace a cap anyway.

[flinx]

thank you all.


12 volts!

[flinx]

Originally Posted by 999999999

I feel this topic is going a bit backwards. Usually, you start out with a project with defined needs, you have a load needing X voltage and Y current and you determine if any of the parts you have are suitable to build a power supply for that. Your power board could be a start to that but what you have on the board is quite basic, it might be more of a hassle to mount a 2nd board than just throw a bridge rectifier, capacitor or two, and a fuse on the load's board, of course if that's reasonably possible.

Also, playing around with LOW voltage circuits is a good way to learn, but when you get to the point where you're ordering parts to build something, the power board and regulator you salvaged are the equivalent of only $1 to $2 worth of parts (depending on your global location), less considering you have to replace a cap anyway.

I tend to do things a bit backwards.

Also I have lots of salvaged boards, etc.

[flinx]

Now a longer response. (I had a cat on my lap).

While I did take an electronics class in high school (30 years ago), I did not get much from the class as they had a TRS-80 model 1, once I saw that, it's all I messed with. The teacher was very laid back (Vietnam vet) that did not run a very strict class. So I did not retain much. I have been doing minor repairs to monitors, AC Adapters, and anything else with bad capacitors. Every once in a while I try something a little more involved than I am used to. I found an old radio and decided to see what I could do with the PS.

Although I had the pinouts for the Voltage regulator I did not want to plug it in without some kind of confirmation. I appreciate your patience and help. Even if I never find a use for this little PS, I think I finally understand Bridge rectifiers, and getting a better understanding of Voltage regulators. I really should have paid more attention in HS instead of blowing up capacitors, Resistors, playing on the computer, and messing with the Short wave radios.

Thank you.